measuring inflation in the modern economy – a micro price ... · globalization (decreased?)...

Measuring inflation in the modern economy – a microprice-setting view

Aviv Nevo1 Arlene Wong 2

1University of Pennsylvania and NBER

2Princeton University and NBER

June 2018

Introduction Measurement Conceptual Conclusion

Introduction

• Trends in advanced economics:• Globalization• (Decreased?) Competition• Increased high FC/low MC industries• Growth of e-commerce and the sharing economy• Price setting behavior: “Big data” and new pricing models

• We have been asked to give an IO/Marketing view on why/how these might changepricing at the micro level

• What are the aggregate implications?• Mostly leave it for the discussion

Nevo and Wong Micro Price-Setting June 2018 2


Overview

• We organize the discussion around two areas:• Measurement

• Substitution bias: shopping behavior• New products and online shopping• Measuring prices online

• Conceptual issues• Cost pass through• Decreased competition• “New” pricing models• Heterogeneity

• To think through some of the issues, one needs a specific question in mind• We will focus (somewhat) on the question of low inflation during the recent recovery



Bottom line

• Not sure we have a bottom line, but to the extent we do• The need for more data• The need for more (collaborative) research



Why is inflation low?

• Inflation has stayed persistently low post the Great Recession• Part of what we have been asked is to give an IO/Marketing view on reasons why this

might be the case• Why do we expect inflation to rise post recession?

• A tightening job market should lead to rising wages• Rising wages (higher cost more generally) should lead to rising prices

• Did wages go up?• Seems like maybe not• We will leave if to labor economists to argue• Could monopsony power be a reason? (Marinescu, Azar, Steinbaum and Taska, 2018)

• Additional reason: rising aggregate demand could bring the production sector closer tocapacity constraints



Other reasons for low inflation

• To an IO economist it is not obvious that even if wages went up that we should see high(measured) inflation

• Measurement of inflation• Imperfect cost pass through• Long term secular trends that are bringing a structural change

• Trends in concentration• Global trade and a more complex supply chain• Changes in how firms set prices

• We provide an initial exploration of some of these topics



The measurement of inflation

• Measuring inflation is notoriously difficult• We will focus on:

• Substitution bias: shopping behavior• New products and online shopping• Measuring prices online



Substitution bias

• “Substitution bias occurs because a fixed market basket fails to reflect the fact thatconsumers substitute when relative prices change” (Boskin committee report)

• Substitution bias can come in many forms:• products• outlets• new goods

• We focus on shopping behavior more generally• Substitution across stores• Substitution across sizes• The use of coupons (or other discounts)• Substitute over time (by stockpiling)



Shopping behavior

• Several papers suggest that shopping behavior changes with economic conditions• Aguiar and Hurst (2007); Griffith et al. (2009); McKenzie and Schargrodsky (2011);

Coibion, Gorodnichenko and Hong (2012); Aguiar, Hurst and Karabarbounis (2013); Beraja,Hurst and Ospina (2015)

• Generally we believe that substitution bias leads to an underestimate of inflation• Nevo and Wong (forthcoming) imply increased shopping behavior, which could lead to an

overestimate of inflation during the recession• In principle inflation could be underestimated during the recovery if behavior reverted



Shopping patterns and the recession: Sales

0.24

0.25

0.26

0.27

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Purchases of Sale Items

Share of total household expenditure

Non-parametric Cubic Linear Spline

Source: US Nielsen Homescan Data



Shopping patterns and the recession: Coupons

0.060

0.070

0.080

0.090

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Purchases with Coupon Use






Shopping patterns and the recession: Generic products

0.08

0.09

0.1

0.11

0.12

0.13

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Purchases of Generic Items






Shopping patterns and the recession: Discount stores

0.28

0.30

0.32

0.34

0.36

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Purchases made at Discount Stores






What impact did this have on inflation?

• Hard to give an exact answer• Back of the envelope computation suggests that true inflation could be as much as 20%

higher than measured inflation



Shopping behavior and the measurement of inflation

• Low inflation post recession• Did shopping behavior revert after the recession? yes!• If so, is the true inflation higher than measured? Maybe

• How about more generally?• Shopping behavior not new• But has there been a change over time?• Maybe: here are a few examples



Product variety• Neiman and Vavra (2018) show a rise in the concentration of household spending• So maybe less product substitution



Online Shopping• On the other hand, online shopping is increasing



Measuring online economic activity

• Online shopping potentially can generate substitution bias• But more basically there is a measurement issue• Hasuman and Leibtag (2004) “CPI Bias from Supercenters: Does the BLS Know that

Wal-Mart Exists?”• How about Amazon? eBay? Uber? AirBnB?

• Are online prices collected?• Are online prices different?• Easy to come up with theory that suggest they would be• Measurement: Goolsbee and Klenow (2018) suggest lower inflation online



Cost pass through

• It might seem intuitive that prices should go up if wages increase• This does need to be the case• Consider a firm f facing a (residual) demand curve Q(p) setting a uniform price p to

maximize

maxp

Π(p) = (p − mc)Q(p) − F

where mc is (constant) marginal cost; and F is fixed cost



Cost pass through

• The FOC is given by

p = mc − Q∂Q/∂p

• It is tempting to conclude that the pass through on marginal cost is 1• In general, this is not true

• The semi-elasticity depends on price (for most demand curves)• The degree of pass through depends on the curvature of the demand curve• In principle, can be more or less than 1• “Typical” demand curves have less or near 1 (e.g. linear 0.5)

• Empirically most products pass through is less than 1



Cost pass through

• Note: imperfect pass through is profit maximizing• It is not the case that firms are “eating the cost”• Is this enough to explain low inflation?• Could work together with labor market forces• How about longer term?



Long term trends: decreased competition

• Much publicity has been given recently to the rise in market power (e.g., De Loecker andEeckhout, 2017)

• Could this be impacting pass through?• Directly tied to decreasing labor share: so less impact on price even holding pass-through

constant• Some believe that pass-through decreases with concentration, which might suggest less

pass-through• Might be US only phenomena

• As an aside• What about welfare? If consumers are paying more and getting better quality are they really

worse off?



More complex supply chain

• An additional long term trend is globalization and a more complex supply chain (Wei andXie, 2018)

• Two implications• Less direct link between local labor and products• More levels of intermediates and therefore even lower(?) pass through



More complex supply chain

5010

015

020

0

1970 1980 1990 2000 2010 2020

PPI CPI

CAN

5010

015

020

025

0

1970 1980 1990 2000 2010 2020

PPI CPI

USA

010

020

030

0

1970 1980 1990 2000 2010 2020

PPI CPI

AUS

5010

015

020

025

0

1970 1980 1990 2000 2010 2020

PPI CPI

NOR

050

100

150

200

1970 1980 1990 2000 2010 2020

PPI CPI

SWE

050

100

150

200

1970 1980 1990 2000 2010 2020

PPI CPI

DNK

PPI and CPI Indexed to 100 in 1985



Changes in how firms set prices

• Up to now we talked about a simple linear uniform price• However, firms are consistently coming up with new forms of pricing• Consider (the not so new idea of) price discrimination

• 1st degree (perfect PD): charge each consumer their WTP• 2nd degree (discriminate based on observable demographics): price to market, student

discounts• 3rd degree (self sorting): coupons, temporary price reductions, product line pricing, bundling

and others• Historically limited amount of 1st and 2nd degree PD ... but with Big Data ...



Price discrimination and inflation

• What does this have to do with low inflation post-recession• In many cases “fancier” pricing can yield lower pass through rates• Takes measurement issues to a whole new level• How do we measure inflation in a world where prices are individualized?

• already today there is heterogeneity in prices levels and inflation• does this matter?



Concluding Comments

• Measurement• Substitution bias: shopping behavior• New products and online shopping• Measuring prices online

• Conceptual issues• Cost pass through• Decreased competition• “New” pricing models• Heterogeneity

• A real need for more data to measure true inflation better• Consistent measurement of online prices• A focus on what consumer pay versus what firms charge

• To understand aggregate application of modern pricing need more collaborative research



References

• Aguiar, Mark, and Erik Hurst. 2007. “Life-cycle Prices and Production.” American Economic Review, 5(3): 1533-1599.• Aguiar, Mark, Erik Hurst, and Loukas Karabarbounis. 2013. “Time Use during the Great Recession.” American Economic

Review, 103(5): 1664-1696. Beraja, Martin, Erik Hurst, and Juan Ospina. 2015. “The Aggregate Implications ofRegional Business Cycles.”

• Coibion, Olivier, Yuriy Gorodnichenko, and Gee Hee Hong. 2012. “The cyclicality of sales, regular and effective prices:Business cycle and policy implications.” American Economic Review, 105(3): 993-1029.

• De Loecker and Eeckhout, 2017. “The Rise of Market Power and the Macroeconomic Implications”. NBER WorkingPaper No. 23687.

• Goolsbee, Austin D. and Peter J. Klenow, 2018. “Internet Rising, Prices Falling: Measuring Inflation in a World ofE-Commerce”, NBER Working Paper No. 24649.



References

• Griffith, Rachel, Ephraim Leibtag, Andrew Leicester, and Aviv Nevo. 2009. “Consumer Shopping Behavior: How MuchDo Consumers Save?” Journal of Economic Perspectives, 23(2): 99-120.

• Hasuman, Jerry and Ephraim Leibtag, 2009. “CPI Bias from Supercenters: Does the BLS Know that Wal-Mart Exists?”in Price Index Concepts and Measurement, Diewert, Greenlees, and Hulten. 2009.

• McKenzie, David and Ernesto Schargrodsky, 2011. “Buying less and shopping more: the use of non-market labor during acrisis”, Economia Journal, 0(20), 1-43.

• Nevo, Aviv and Arlene Wong (2018), “The Elasticity of Substitution Between Time and Market Goods: Evidence fromthe Great Recession”, International Economic Review, forthcoming.

• Wei Shang-Jin and Yinxi Xie, 2018, “On the Divergence between CPI and PPI as Inflation Gauges: The Role of SupplyChains”, NBER Working Paper No. 24319.


measuring inflation in the modern economy – a micro price ... · globalization (decreased?)...

Documents